Mark Wilson's Blog

Computers and heat

For a few months now I've been having overheating problems with my computer. Random restarting in the middle of your work can be very annoying, let me tell you. However, at least the cause was relatively clear: Way back when, I purchased an el cheapo microATX system with no fans, and slowly upgraded, so lately there's been a bigger, heat-generating graphics card; two hard drives sitting right on top of each other; two optical drives sitting on top of one another; and a power supply, less than an inch from the optical drives, appearing to take up half the case.

I had an awesome high-tech setup for dealing with this: Side came off the case, desktop fan sat next to the computer and blew in.

A couple days ago I bought a new mid-tower case, armed with two honkin' 120-mm fans, and moved everything over into it. The result of this grand planning was that the computer continued with its random restarts, particularly under load. In bafflement, I took a closer look at everything and discovered that my CPU heatsink was caked in dust under its fan, probably as a result of three years in the dorms compounded with several months of open-case operation.

However, blasting the heatsink (and everything else in sight) completely free of the insidious dust bunnies doesn't seem to have totally done the trick, either. I suffered another restart shortly after. It seems a touch more stable now, probably a half-hour later, but I'm less than confident here. Wedging the case's temperature probe under the CPU as best I can (I can't seem to get the heatsink off) gets me a report of a running temp around 46 degrees Celsius, which is well under the common 60-degree worry mark and a mile away from the manufacturer's 85-degree maximum operating temperature.

I'm about out of ideas here. Despite my upgrades, this is at heart a three-year-old system, nothing outrageously powerful. Nothing in the computer is overclocked. I can't think of a single reason for this system to be so twitchy.